A Mouse PRMT1 Null Allele Defines an Essential Role for Arginine Methylation in Genome Maintenance and Cell Proliferation

Abstract

Protein arginine methyltransferase 1 (PRMT1) is the major enzyme that generates monomethylarginine and asymmetrical dimethylarginine. We report here a conditional null allele of PRMT1 in mice and that the loss of PRMT1 expression leads to embryonic lethality. Using the Cre/lox-conditional system, we show that the loss of PRMT1 in mouse embryonic fibroblasts (MEFs) leads to the loss of arginine methylation of substrates harboring a glycine-arginine rich motif, including Sam68 and MRE11. The loss of PRMT1 in MEFs leads to spontaneous DNA damage, cell cycle progression delay, checkpoint defects, aneuploidy, and polyploidy. We show using a 4-hydroxytamoxifen-inducible Cre that the loss of PRMT1 in MEFs leads to a higher incidence of chromosome losses, gains, structural rearrangements, and polyploidy, as documented by spectral karyotyping. Using PRMT1 small interfering RNA in U2OS cells, we further show that PRMT1-deficient cells are hypersensitive to the DNA damaging agent etoposide and exhibit a defect in the recruitment of the homologous recombination RAD51 recombinase to DNA damage foci. Taken together, these data show that PRMT1 is required for genome integrity and cell proliferation. Our findings also suggest that arginine methylation by PRMT1 is a key posttranslational modification in the DNA damage response pathway in proliferating mammalian cells.

Methylation of arginine residues is one of many posttranslational modifications of eukaryotic proteins (7). Arginine methylation is catalyzed by a family of enzymes named protein arginine methyltransferases (PRMTs). PRMTs catalyze the direct transfer of a methyl group from S-adenosyl-l-methionine to one or two of the guanidino nitrogen atoms in arginine (7). In higher eukaryotes, there are 11 PRMTs classified into two groups according to their substrate specificity and reaction products (6). Type I enzymes, including PRMT1 (39), PRMT3 (54), CARM1 (16), PRMT6 (25), and PRMT8 (37), catalyze the formation of N^G-monomethylarginine and asymmetric N^G,N^G-dimethylarginine, while type II enzymes, including PRMT5 (45), and PRMT7 (41), catalyze the formation of N^G-monomethylarginine and symmetric N^G,N′^G-dimethylarginine. PRMT2 (49) has no detectable activity, and the activity of PRMT9 has not been determined (6). FBXO10 and FBXO11 were proposed as PRMT10 and PRMT11 (19).

The PRMT1 and PRMT3 genes have been targeted in mouse embryonic stem (ES) cells using gene-trapping strategies (44, 53). The targeted alleles in both cases result in hypomorphic alleles with ∼5% residual PRMT1 and PRMT3 expression, respectively (20, 53). Mice homozygous for the PRMT1 gene-trap hypomorphic allele die at around embryonic day 6.5 (E6.5). ES cells were isolated that are homozygous for the PRMT1 hypomorphic allele, and these cells harbor numerous hypomethylated proteins (44), including Sam68 (20), MRE11 (12), histone H4 (52, 61), and hnRNPK (43). PRMT3 null mice have retarded growth during gestation but develop normally thereafter. Mouse embryonic fibroblasts (MEFs) derived from these PRMT3 cells harbor hypomethylated ribosomal protein rpS2 (53).

PRMT2 and CARM1 null mice have been generated by gene targeting using homologous recombination (HR). PRMT2 null mice are viable with no gross abnormalities (68). However, the PRMT2^−/− MEFs have elevated NF-κB activity and decreased susceptibility to apoptosis (27). PRMT2^−/− MEFs also have an earlier S phase entry by bromo-2′-deoxyuridine (BrdU) staining, but the growth profiles are similar to those of wild-type MEFs (68). CARM1^−/− mice survive to birth but die perinatally (66). CARM1^−/− mice have a defect in thymocyte maturation at an early progenitor stage (35) and an adipogenesis defect (65). CARM1 serves as a coactivator for numerous transcription factors, including nuclear receptors, p53, NF-κB and MEF2C (6).

PRMT1 is the predominant type I PRMT in mammalian cells, responsible for at least 85% of all arginine methylation reactions in human cells (55). Saccharomyces cerevisiae without the PRMT1 homolog (hmt1/Rmt1) are viable, mislocalize cellular proteins (50) and harbor defects in maintaining silent chromatin (4, 69). PRMT1 catalyzes substrate dimethylation in a partially processive manner (42) and oligomerizes into ring-like structures (70). A large number of PRMT1 substrates are known, and its preferred methylation sites are arginines that lie within glycine- and arginine-rich (GAR) sequences that include multiple arginines in RGG or RXR contexts (7). As such, antibodies have been generated that recognize methylated GAR regions and are good indicators of PRMT1 activity within the cell (10). PRMT1 substrates that lack a GAR motif have been identified including, PGCα (56), the estrogen receptor (38), and FOXO transcription factors (67). As reflected by the diversity of its substrates, PRMT1 is implicated in the regulation of a plethora of cellular processes. For example, PRMT1 is implicated in the transcriptional coactivation of nuclear hormone receptors because it methylates histone H4 (52, 61) and thereby facilitates histone acetylation and chromatin remodeling (3). In addition, PRMT1 methylates the RNA-binding protein Sam68 (20) and the DNA damage response (DDR) proteins MRE11 (11, 21) and 53BP1 (1, 13). MRE11 forms a complex with RAD50 and NBS1 and is referred to as the MRN complex (48). More recently, PRMT1 has been shown to regulate the cytoplasmic signaling function of the estrogen receptor α (38). In addition to its numerous cellular function, the PRMT1 activity is dysregulated in cancer (7). PRMT1 is aberrantly expressed in prostate cancer and likely contributes to the proliferative capacity of prostate cancer cells through its ability to act as a transcriptional coactivator for the androgen receptor (7). Moreover, the knockdown of PRMT1, or its substrate Sam68, suppressed mixed lineage leukemia-mediated transformation (18).

In the present study, we report the generation of the first PRMT1 null allele in mice. We show that the MEFs lacking PRMT1 exhibit spontaneous DNA damage, cell cycle delays, checkpoint activation defects after DNA damage, polyploidy, and chromosome instability. Moreover, PRMT1 knockdown U2OS cells are hypersensitive to etoposide and have an impaired ability to recruit the RAD51 recombinase to DNA damage sites. These findings demonstrate that arginine methylation by PRMT1 plays a key role in genome maintenance and the DDR pathway.

MATERIALS AND METHODS

Reagents and antibodies.

BrdU, 4-hydroxytamoxifen (OHT), and Polybrene were purchased from Sigma (St. Louis, MO). Mouse anti-γH2AX monoclonal antibody, rabbit anti-phospho-histone H3-Ser10, anti-Sam68, anti-PRMT1, anti-PRMT5, anti-PRMT7, and ASYM24 antibodies were obtained from Millipore (Billerica, MA). Rabbit anti-MRE11 and anti-53BP1 antibodies were from Novus Biologicals (Littleton, CO). Rabbit anti-RAD51 antibody was from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-CARM1 was generated as described previously (35). The anti-PRMT6 antibody was purchased from IMGENEX (San Diego, CA). ASYM25b antibody was generated in rabbits using asymmetrically dimethylated GAR motif as described previously (10). Propidium iodide (PI) and Alexa Fluor 488-conjugated goat anti-rabbit and anti-mouse immunoglobulin G (IgG) antibodies were from Invitrogen (Carlsbad, CA). Fluorescein isothiocyanate (FITC)-conjugated mouse IgG antibody was obtained from Sigma. FITC-conjugated mouse anti-BrdU monoclonal antibody was from BD Biosciences, Pharmingen (San Diego, CA). Immunoprecipitations and immunoblotting were performed as previously described (9).

Cell culture, retrovirus infection, and transfection.

We constructed the PRMT1 conditional allele in mice essentially as described previously (17, 33). The sequenced plasmids were linearized and electroporated into 129/Sv ES cells. Three independent ES cell clones with homologous integration at the targeting site were injected into C57BL/6J blastocysts, and chimeras were obtained. These chimeras were subsequently crossed with C57BL/6J females, and the heterozygous mice with successful germ line transmission of the targeted allele were crossed with C57BL/6J mice expressing Flp recombinase to remove the neomycin resistance cassette, resulting in a floxed allele (PRMT^FL/+). To disrupt PRMT1, heterozygous PRMT1^FL/+ mice were crossed with Zp3-Cre mice, resulting in a heterozygous deletion (PRMT1^+/−) of the methyltransferase domain for PRMT1. We bred the PRMT1^FL/+ mice with the PRMT1^+/− mice to generate PRMT1^FL/− MEFs from E14.5 embryos. Spontaneously immortalized MEFs were created according to standard 3T3 protocol. U2OS cells were obtained from American Type Culture Collection (Manassas, VA). All cells were grown in Dulbecco modified Eagle medium containing 10% bovine calf serum.

Retroviral vectors encoding green fluorescent protein (GFP), GFP-CRE fusion protein, or CRE and hygromycin marker (hygro-Cre) were from Guy Sauvageau (Institut de Recherche en Immunologie et en Cancérologie, Montréal, Québec, Canada). Viral particles were produced using an ecotropic packaging system and used for infection of primary or immortalized MEFs as described previously (40).

Immortalized MEFs were transfected by a plasmid DNA encoding estrogen receptor-CRE fusion protein (ER-CRE) and blasticidin selection marker (pCAG-IRESblast-ER-CRE) (17) using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Stable cell lines were selected with 3 μg of blasticidin/ml, and a few single clones were collected. The cells were treated with the estrogen receptor ligand, OHT, to induce relocalization of the ER-CRE fusion protein. OHT was removed from the culture medium 2 days before performing a phenotype analysis. Human osteosarcoma U2OS cells were transfected by small interfering RNAs (siRNAs) using Lipofectamine RNAi MAX (Invitrogen) according to the manufacturer's instructions with a final concentration of 20 nM siRNA. The siRNAs were purchased from Dharmacon (Lafayette, CO). The target sequence of PRMT1 siRNA was 5′-CGTCAAAGCCAACAAGTTA-3′. A GFP siRNA served as the control with the sequence 5′-AACACUUGUCACUACUUUCUCUU-3′.

Clonogenic assay.

A total of 200 to 400 cells were plated on 10-cm dishes and treated with various concentrations of reagents as indicated in the figure legends in duplicate 24 h after plating. For the treatment of U2OS cells with etoposide, the cells were washed twice with phosphate-buffered saline (PBS) after the treatment. The colonies containing approximately 20 or more cells were counted at 10, 12, and 14 days after treatment. For the treatment of immortalized MEFs with blasticidin and OHT, cells were maintained in the medium containing the desired reagents until staining. The cells were allowed to grow for 12 days, and the colonies were fixed and stained with 0.1% crystal violet for 30 min. The stained colonies were counted. The surviving fraction was determined by dividing the average number of colonies for each treatment by the average number of colonies in the control dishes.

Ionizing radiation (IR) treatment.

Cells were irradiated at 20 to 24 h after plating at room temperature by using a Theratron T-780 cobalt unit located in the Department of Radiation Oncology at the Jewish General Hospital (Montreal, Quebec, Canada). Doses ranging from 2 to 10 Gy were delivered at a dose rate 0.66 Gy/min. The cells were returned to an incubator after the irradiations and maintained at 37°C for further analysis.

Cell cycle analysis.

For all flow cytometry experiments, both cells growing on the surface of the dishes and in the culture medium were harvested, fixed with 75% ethanol, and stored at −20°C for less than 1 week before staining and analysis. All flow cytometry measurements were performed by using FACSCalibur flow cytometer (BD Biosciences). The data were analyzed by using BD CellQuest Pro software. For measurement of BrdU incorporation, cells were incubated with 10 μM BrdU for 45 min before they were harvested. Cells were fixed, washed once with PBS, and treated with 2 M HCl for 20 min. Cells were then neutralized with 0.1 M sodium borate (pH 9.0), washed once with PBS containing 1% bovine serum albumin, and stained with FITC-conjugated anti-BrdU antibody diluted in 100 μl of PBS containing 1% bovine serum albumin and 0.1% Tween 20 (PBS-T) for 20 min. After two washes with PBS-T, cells were resuspended in PBS containing PI at 10 μg/ml and RNase A at 0.2 mg/ml. After 20 to 60 min of incubation, the cells were subjected to flow cytometry analysis. For measurement of phosphorylated histone H3, fixed cells were first stained with anti-phospho-histone H3 (Ser 10) rabbit antibody, and then with FITC-conjugated goat anti-rabbit IgG (Invitrogen) after being washed with dilution buffer (1% fetal bovine serum and 0.1% Triton X-100 in PBS) as described previously (58). Cells were then counterstained with PI. More than 10,000 cells per condition were analyzed by flow cytometry.

Immunofluorescence.

Cells growing on glass coverslips were washed with PBS twice and fixed with 4% paraformaldehyde at room temperature for 10 min. The cells were then permeabilized in permeabilization solution (0.5% Triton X-100 in PBS) for 15 min. After three washes with PBS, cells were blocked with 10% goat serum in PBS and stained with mouse anti-γH2AX (1:1,000), rabbit anti-53BP1 (1:200), or rabbit anti-RAD51 (1:20) antibodies diluted in PBS containing 5% goat serum and 0.1% Triton X-100. After three washes, the cells were then stained with Alexa Fluor 488-conjugated goat anti-rabbit and Alexa Fluor 568-conjugated goat anti-mouse secondary antibodies. DNA was counterstained with DAPI (4′,6′-diamidino-2-phenylindole) after three washes with PBS, and coverslips were mounted with Immuno-Mount purchased from Thermo Scientific. Images were taken by using a Zeiss M1 fluorescence microscope.

SKY analysis.

The PRMT1^FL/− MEFs were left untreated (−OHT) or treated for 4 days with OHT and incubated for another 2 days without OHT (+OHT, 6 days) and analyzed by spectral karyotyping (SKY) at the Banque de Cellules Leucémiques du Québec (Centre de Recherche, Hôpital Maisonneuve-Rosemont, Montreal, Quebec, Canada). Slide pretreatment, hybridization with the SkyPaint mouse probes, and detection were performed according to the protocol provided by Applied Spectral Imaging (ASI) with minor modifications. Spectral images were acquired with a SpectraCube system (ASI) mounted on a Zeiss Axioplan II microscope and analyzed using SkyView version 1.6.1 software (ASI). Twenty-four and twenty-eight metaphases were analyzed for the −OHT- and +OHT-treated PRMT1^FL/− MEFs.

RESULTS

Generation of PRMT1 conditional and null alleles in mice.

Using the Cre/loxP recombination system, we generated a PRMT1 conditional allele (PRMT1^FL) that consists of exons 4 and 5 flanked by loxP sites. Cre recombinase-mediated deletion of the “floxed” region will remove part of the methyltransferase domain, including the S-adenosyl-l-methionine-binding site, and create a frameshift, thereby resulting in a functionally null allele (PRMT1⁻) (Fig. 1A). PRMT1^FL/+, PRMT1^FL/FL, and PRMT1^+/− mice were grossly normal and fertile, whereas PRMT1^−/− embryos did not survive to 7.5 days postcoitum, the earliest time point examined. The role of PRMT1 in embryonic development and adult tissues is under investigation. In the present study, we addressed the cellular function of PRMT1 using MEFs.

FIG. 1.

Generation of a PRMT1 conditional allele in mice and PRMT1-deficient MEFs. (A) Schematic representation of the wild-type (PRMT1 locus), floxed (2 loxP; PRMT1^FL), and null (1 loxP; PRMT1⁻) PRMT1 alleles. The exons are the black boxes, and the line represents introns not drawn to scale. The black triangles denote loxP sites, and the small arrows denote the primers (named CT#) used for PCR analysis. The expected size of the CT-497/498 DNA fragment for the wild-type allele is 245 bp, while the size of the DNA fragment for the 2loxP allele is 358 bp. The expected size of the CT-296/498 DNA fragment after CRE excision is 294 bp. (B) Primary MEFs deficient for PRMT1 were generated by infecting cells with a hygro-CRE retrovirus. Genomic DNA from hygromycin-selected cells was analyzed by PCR, and the DNA fragments were visualized on an ethidium bromide-stained agarose gel. M denotes molecular mass markers of the 1-kb ladder (Invitrogen, Inc.). (C) PRMT1 MEFs of the indicated genotype infected with hygro-Cre or not were lysed, and the total cellular proteins were analyzed by immunoblotting with anti-PRMT1 and anti-α-tubulin antibodies, as a loading control. The migration of PRMT1 and α-tubulin is shown on the right with arrows, and the migration of the molecular mass markers is shown on the left in kilodaltons. (D) PRMT1^FL/− MEFs were immortalized and stably transfected with a plasmid encoding the estrogen receptor-CRE fusion protein (ER-CRE). The cells, termed PRMT1^{FL/−;CreERT} MEFs or control PRMT1^+/+;CreERT MEFs, were then treated with OHT for 0, 2, 4, and 6 days, and genomic DNA was isolated and analyzed as in panel B. (E) Total cellular proteins from mock-treated (−) or OHT-treated (+) PRMT1^{FL/−;CreERT} and PRMT1^+/+;CreERT MEFs were analyzed by immunoblotting as described for panel C.

PRMT1-deficient MEFs.

We isolated MEFs from 14.5-day-postcoitum embryos and generated PRMT1^+/+ and PRMT1^FL/− primary MEFs. To disrupt PRMT1, we infected these primary MEFs with hygromycin-resistant retroviruses that express Cre recombinase (hygro-Cre). The Cre recombinase catalyzed the deletion of the exons between the two loxP sites of PRMT1^FL/− allele, resulting in PRMT1-deficient MEFs. PCR amplification of a DNA fragment from genomic DNA isolated from PRMT1^FL/− showed that the presence of Cre led to the loss of the DNA fragment for 2loxP and the gain of the 1loxP DNA fragment (Fig. 1B, lanes 4 and 6). Moreover, we stably transfected spontaneously immortalized PRMT1^FL/− MEFs with a plasmid encoding the estrogen receptor-CRE fusion protein (PRMT1^{FL/−;CreERT}). The addition of OHT for 2, 4, and 6 days led to loss of the 2loxP DNA fragment (Fig. 1D, lanes 6, 7, and 8 [upper panel]), while the 1loxP DNA fragment was observed (Fig. 1D, lanes 5 to 8 [lower panel]). Immunoblotting total cellular extracts from PRMT1^FL/− MEFs infected with hygro-Cre retroviruses and PRMT1^{FL/−;CreERT} MEFs treated with OHT showed a complete loss of PRMT1 expression, including the slower-migrating spliced isoform of 48 kDa (Fig. 1C and E) (20). The deletion of exons 4 and 5 is expected to lead to a frameshift and, indeed, we did not observe a truncated protein (Fig. 1C and E). These findings confirm that we have generated a PRMT1 null allele.

The loss of PRMT1 in MEFs leads to the hypomethylation of cellular proteins, including Sam68 and MRE11.

To determine whether PRMT1 is functionally deleted, we immunoblotted total cellular extracts of PRMT1^+/+ and PRMT1^FL/− MEFs with two methylarginine-specific antibodies that recognize proteins with methylated GAR motifs (10). The infection of PRMT1^FL/− MEFs with hygro-Cre resulted in hypomethylation of many cellular proteins, as detected with ASYM24 (Fig. 2A) and ASYM25B (Fig. 2B). This hypomethylation was not observed in PRMT1^+/+ MEFs infected with Cre (Fig. 2A and B). To further confirm the deficiency of PRMT1 function, we immunoprecipitated previously defined PRMT1 substrates, including Sam68 and MRE11, and examined their methylation status. PRMT1^FL/− MEFs left untreated or infected with a hygro-CRE retrovirus were immunoprecipitated with anti-Sam68 antibodies and immunoblotted with either anti-Sam68 as control or anti-ASYM24 antibodies to monitor its methylation. The hypomethylation of Sam68 was clearly visible, since the immunoprecipitated Sam68 was not recognized by ASYM24 in the Cre-transduced cells (Fig. 2C). We next utilized PRMT1^{FL/−;CreERT} MEFs treated with OHT for 6 days (+OHT) or left untreated (−OHT). Note that 6 days of OHT treatment means the first 4 days with OHT and the subsequent 2 days without OHT, as described in Materials and Methods. Cell lysates were prepared from OHT-treated cells and immunoprecipitated with anti-MRE11 antibodies. The bound proteins were analyzed by immunoblotting with either anti-MRE11 antibodies as a control or anti-ASYM25b antibodies to monitor its methylation. ASYM25b weakly recognized immunoprecipitated MRE11 from the OHT-treated cells, demonstrating that it is hypomethylated in these cells (Fig. 2D). These results confirm that PRMT1 was functionally deleted from the PRMT1-deficient MEFs, since two of its well-known substrates are hypomethylated. We next immunoblotted with antibodies against other PRMTs, including PRMT3, CARM1 (also known as PRMT4), PRMT5, PRMT6, and PRMT7, to confirm that the loss of PRMT1 does not affect the expression of the other PRMTs. Indeed, the expression of the other PRMTs was not altered with the loss of PRMT1 (Fig. 2E). These findings suggest that the hypomethylation of cellular proteins in PRMT1-deficient MEFs is caused by the loss of PRMT1 and not the simultaneous decrease of another PRMT.

FIG. 2.

Hypomethylation of cellular proteins in PRMT1-deficient MEFs. (A and B) PRMT1^+/+ and PRMT1^FL/− MEFs were left untreated (-CRE) or infected with the hygro-Cre retrovirus and selected with hygromycin (+CRE). Cell lysates were immunoblotted with anti-asymmetrical dimethylarginine antibodies ASYM24 (A) or ASYM25b (B) and anti-α-tubulin antibodies to control for equivalent loading. The migration of the molecular mass markers is shown on the left in kilodaltons. (C) PRMT1^FL/− MEFs were left untreated (-CRE) or infected with hygro-Cre retroviruses (+CRE) and selected with hygromycin. Whole-cell lysates were immunoprecipitated with anti-Sam68 antibodies. The bound proteins were separated by SDS-polyacrylamide gel electrophoresis and immunoblotted with anti-Sam68 (left panel) and ASYM24 (right panel) antibodies as indicated. The migration of Sam68 and the heavy chain of IgG is shown. (D) PRMT1^{FL/−;CreERT} MEFs were treated with (+OHT) or left untreated (-OHT) for 6 days. Whole-cell lysates were immunoprecipitated with anti-MRE11 antibodies, and the bound proteins were separated by SDS-polyacrylamide gel electrophoresis and immunoblotted with anti-MRE11 and ASYM25b antibodies as indicated. The migration of MRE11 is shown. (E) PRMT1^FL/− MEFs were left untreated (-CRE) or infected with hygro-Cre retroviruses (+CRE) and selected with hygromycin. Whole-cell lysates were immunoblotted with the indicated PRMT antibodies or the anti-tubulin antibody to visualize equivalent loading.

The loss of PRMT1 results in cell growth arrest.

We initially compared the growth characteristics of primary PRMT1^+/+, PRMT1^FL/+, and PRMT1^FL/− MEFs. We observed that removal of one allele of PRMT1 did not significantly affect cell proliferation (Fig. 3A). To study the function of PRMT1 in cell proliferation, we infected PRMT1^FL/− MEFs with retroviruses that express GFP alone or GFP-CRE and obtained ∼70% of the cells to be GFP positive 5 days after infection, as assessed by fluorescence-activated cell sorting (Fig. 3B). However, at 12 days postinfection very few GFP-positive cells remained (∼15%) in the Retro-GFP-Cre-infected PRMT1^FL/− MEFs (Fig. 3B). These findings suggest that PRMT1 deficiency causes cell death or cell cycle arrest, and thus the cells that were not transfected with CRE overpopulate the PRMT1-deficient slower-growing cells. The GFP-positive cells at day 5 were normalized to 100%, and they were monitored for 2 weeks postinfection for GFP expression and expressed as a percentage. The presence of the Cre recombinase led to ∼60% and ∼20% survivals of GFP-positive PRMT1^FL/− MEFs at 9 and 14 days after retroviral infection, respectively (Fig. 3C). PRMT1^FL/− MEFs that were infected with GFP alone survived and maintained their GFP expression (Fig. 3C). These results show that the survival of PRMT1-deficient MEFs may be compromised, suggesting that PRMT1 may be required for the viability of MEFs. To further confirm the requirement of PRMT1 for cell survival and/or proliferation, we next performed a colony formation assay using PRMT1^{FL/−;CreERT} MEFs treated with blasticidin, the antibiotic selection of the CreERT plasmid, and OHT. The generation of PRMT1-deficient cells with ER-CRE (+ER−CRE; blasticidin) and OHT led to no colonies (Fig. 3D). As controls, colonies were obtained with PRMT1^FL/− treated with OHT and PRMT1^{FL/−;CreERT} without OHT treatment. These results show that PRMT1-deficient MEFs die or are growth arrested.

FIG. 3.

PRMT1 is essential for the viability of MEFs. (A) The indicated primary PRMT1 MEFs (P3) were plated at cell density of 5 × 10⁵ cells per 10-cm tissue culture dish. Every two or 3 days, the cells were treated with trypsin and counted with a Beckman-Coulter counter, and a quarter of the cells were replated for the next counting (n = 4). (B) PRMT1^FL/− MEFs were left uninfected (no infection) or infected with retroviruses encoding GFP (GFP) and a GFP-CRE (GFP-CRE) fusion protein, respectively. The GFP-positive cells were identified and counted by flow cytometry at 5 and 12 days postinfection. The y axis denotes the cell number, and the x axis denotes the GFP fluorescence. (C) The quantification of panel B expressed as a percentage of GFP-positive cells normalized to 100% with day 5. (D) Four hundred PRMT1^FL/− (-ER-CRE) and PRMT1^{FL/−;CreERT} (+ER-CRE) MEFs were seeded on 10-cm tissue culture dishes and maintained in regular medium (Dulbecco modified Eagle medium:10% serum) containing different combinations of blasticidin and OHT as indicated. Twelve days later, the cells were fixed, and cell colonies were stained with crystal violet.

PRMT1-deficient MEFs have >4N DNA content and accumulate at the G₂/M phase.

To identify the cellular defect(s) of PRMT1-deficient MEFs, we first examined PRMT1 null MEFs for cell cycle defects. We observed that the number of PRMT1^{FL/−;CreERT} MEFs with >4N DNA content gradually increased up to ∼12% after 8 days of OHT treatment (Fig. 4A), and this corresponded to the loss of PRMT1 expression, as detected by immunoblotting (Fig. 4B). The OHT treatment did not induce the accumulation of PRMT1^+/+ MEFs at the G₂/M phase, nor did we observe a DNA content >4 N in these cells (Fig. 4C). PRMT1-deficient MEFs did not enter aberrant apoptosis 8 days after OHT treatment, since no significant sub-G₁ peak was observed (Fig. 4A). The absence of substantial cell death and the presence of polyploidy suggest that the loss of PRMT1 results in cells that are growth arrested and polyploid.

FIG. 4.

PRMT1-deficient MEFs exhibit >4N DNA content and accumulation of cells at G₂/M phase. (A) PRMT1^{FL/−;CreERT} MEFs were treated with OHT for the indicated days and the cells were stained with PI and analyzed by cell sorting. The percentage of cells containing >4N DNA content is indicated. The experiments were performed three times, and a representative profile is shown. (B) Cell extracts from OHT-treated PRMT1^{FL/−;CreERT} MEFs were immunoblotted with anti-PRMT1 and antitubulin antibodies to confirm the loss of PRMT1. (C) PRMT1^+/+;CreERT MEFs were treated with OHT for the indicated days, and the cells were stained with PI and analyzed by cell sorting. The percentage of cells containing >4N DNA content is indicated.

PRMT1-deficient MEFs exhibit S-phase reduction and cell cycle delay.

To further study the effects of PRMT1 deletion on cell cycle progression, we examined the progression through the S phase using a pulse-chase analysis with BrdU. We treated PRMT1^{FL/−;CreERT} MEFs with OHT for 6 and 10 days to generate PRMT1-deficient MEFs. These cells were compared to untreated PRMT1^{FL/−;CreERT} MEFs (−OHT). The cells were pulsed with BrdU for 45 min and subsequently chased for 2, 4, 6, and 9 h. In the absence of OHT, at 0 h after BrdU incorporation ∼47% of the PRMT1^{FL/−;CreERT} MEFs were BrdU positive, and this number increased to ∼66% at 9 h after BrdU labeling, a finding consistent with cells cycling (Fig. 5A). In contrast, we observed that PRMT1^{FL/−;CreERT} MEFs treated with OHT for 6 and 10 days had a significant decrease in the number of cells in S phase compared to PRMT1^{FL/−;CreERT} MEFs without OHT treatment, and this number decreased slightly after BrdU labeling (Fig. 5A). We next examined the ability of the BrdU-positive cells to progress into mitosis and back into the G₀/G₁ phase of the cell cycle. The majority of the BrdU-positive PRMT1^{FL/−;CreERT} MEFs without OHT (−OHT) progressed within 4 h to the G₂/M phase of the cell cycle (4N) (Fig. 5B, −OHT), and by 6 h they reached the G₀/G₁ phase (2N) (Fig. 5B, −OHT). In comparison, it took 6 h for BrdU-positive PRMT1^{FL/−;CreERT} MEFs with OHT (OHT 6 and 10 days) to progress to the G₂/M phase of the cell cycle. These findings show that PRMT1-deficient MEFs are delayed in cell cycle progression.

FIG. 5.

PRMT1-deficient MEFs exhibit cell cycle delays. (A) PRMT1^{FL/−;CreERT} MEFs were left untreated (-OHT) or treated with OHT (OHT 6 and 10 days) to generate PRMT1-deficient MEFs. The cells growing in log phase were incubated with 10 μM BrdU for 45 min, washed twice with complete medium, and then cultured in the absence of BrdU for the indicated times. The experiments were performed twice, and a representative profile is shown. (B) Histogram representation of BrdU-positive cells. The cells from the boxed area of panel A were represented as a histogram. The 2N and 4N DNA is indicated.

PRMT1^−/− MEFs exhibit spontaneous DNA damage.

The polyploidy and the delayed cell cycle progression suggested that the PRMT1^−/− MEFs display a phenotype reminiscent of defects in the HR pathway, which also display spontaneous DNA damage, sensitivity to DNA-damaging agents, and checkpoint defects (2). In proliferating cells, DNA double-strand breaks (DSBs) occur mainly during DNA replication and an early marker of DNA damage is the phosphorylation of serine 139 of the histone H2AX variant (γH2AX) (47). To examine whether the PRMT1^−/− MEFs have spontaneous DNA damage, we counted the number of γH2AX and 53BP1 foci in PRMT1^{FL/−;CreERT} MEFs with or without 6 days of OHT treatment (Fig. 6A). We observed that ∼30% of the PRMT1^{FL/−;CreERT} MEFs treated with OHT had >5 γH2AX and >5 53BP1 foci compared to ∼10% of the non-OHT treated MEFs, as visualized by indirect immunofluorescence (Fig. 6B). This increase in γH2AX was also observed by immunoblotting after 4 and 8 days OHT treatment (Fig. 6C). These findings show that PRMT1^−/− MEFs contain increased spontaneous DNA damage.

FIG. 6.

PRMT1-deficient MEFs exhibit spontaneous DNA damage. (A) PRMT1^{FL/−;CreERT} MEFs were left untreated (-OHT) or treated with OHT for 6 days (+OHT). The cells were then visualized by indirect immunofluorescence microscopy using anti-γH2AX and anti-53BP1 antibodies. DAPI staining was used to stain the nucleus. (B) The percentage of cells with >5 foci was counted from three independent experiments. The statistical significance was assessed using the Student t test. (C) PRMT1^{FL/−;CreERT} MEFs treated for 0, 4, and 8 days with OHT were lysed and analyzed by immunoblotting with anti-γH2AX and antitubulin antibodies.

Loss of PRMT1 leads to genomic instability in MEFs.

OHT-treated and nontreated PRMT1^{FL/−;CreERT} MEFs were analyzed by SKY. For the definition of the chromosomal abnormalities, the inverted-DAPI banding and spectral images were compared to the SKY-painted chromosomes of the same cell. The most common chromosome number in both cell lines was hypotetraploid and common clonal aberrations, including losses of chromosomes and structural rearrangements that were identified in the two cell lines (Table 1). Representative metaphases are shown in Fig. 7 (see also Fig. S1, S2, and S3 in the supplemental material). We noted that the +OHT cell line had frequent additional aberrations, including a higher incidence of chromosome losses and gains with cells >90 chromosomes (Fig. 7A). The loss of PRMT1 also resulted in the presence of several cells with unique chromosome translocations (Table 1 and Fig. 7B; also see Fig. S3 in the supplemental material). In addition, we noted metaphases with dicentric chromosomes which may suggest end-to-end fusions and with centromeric fusions (Fig. 7B). The presence of 90 to151 chromosomes only in metaphases of the +OHT cell line was observed in 5 of 28 cells, and this shows that the loss of PRMT1 leads to polyploidy. These findings are consistent with cell cycle analysis of Fig. 4 and suggest that the loss of PRMT1 leads to genomic instability.

TABLE 1.

SKY analysis of OHT− and OHT+ MEFs^a

Clonal abnormality	OHT− cell line	OHT+ cell line
Modal no.	69∼73 (4n-) (24 metaphases/24 analyzed)	65∼79 (4n-) (17 metaphases/28 analyzed)
Chromosome no.
1		-1 (7 cells)/+Del(1) (2 cells)/+Del(1)x2 (2 cells)
2	-2 (24 cells)	-2 (17 cells)
3		-3 (4 cells)/Del(3) (4 cells)
4	+Del(4) (3 cells)	-4 (4 cells)/Del(4) (3 cells)
5		-5 (4 cells)/Del(5) (3 cells)
6	Del(6) (5 cells)/Del(6)x2 (17 cells)	Del(6) (5 cells)/Del(6)x2 (12 cells)
7	-7 (19 cells)/-7,-7 (5 cells)	-7 (15 cells)/Del(7)x2 (2 cells)
8	-8 (3 cells)/-8,+T(8;2;6;18) (19 cells)	-8,+T(8;2;6;18) (16 cells)
9	-9 (8 cells)/T(9;18) (4 cells)/T(9;18)x2 (17 cells)	-9 (14 cells)/T(9;18) (4 cells)/T(9;18)x2 (12 cells)
10		-10,-10 (7 cells)
11	T(11;6;11) (23 cells)	-11 (5 cells)/T(11;6;11) (10 cells)
12		-12 (6 cells)
13	-13 (20 cells)	-13 (15 cells)
14		-14 (5 cells)
15	-15 (13 cells)	-15 (12 cells)
16	-16 (24 cells)/Dic(16;16) (21 cells)	-16 (12 cells)/-16,-16 (5 cells)/Dic(16;16) (15 cells)
17	-17 (6 cells)	-17 (6 cells)
18	-18 (20 cells)/T(18;6;8;9) (5 cells)	-18 (7 cells)/-18,-18 (7 cells)/T(18;6;8;9) (12 cells)
19	-19 (20 cells)	-19 (13 cells)/-19,-19 (3 cells)
Single cell abnormalities
Translocations/other aberrations	Isochromosome 4q, T(6;18), T(8;17), T(8;19)	Dic(1;2), Dic(4;5), Dic(5;7), Dic(11;19), T(1;3), T(1;13), T(1;15), T(3;11), T(6;18), T(7;12), T(9;18;9), T(10;?), T(10;2), T(12;7), T(14;15), T(15;10), +10, +11,+ 14, Del(13), Del(18)
End-to-end fusions (?TAs)	None	(1;4), (1;9), (2;2), (3;3), (5;10), (8;9), (18;18)
Other numerical aberrations	None	42∼59 chromosomes (6 metaphases/28 analyzed); 90∼151 chromosomes (5 metaphases/28 analyzed)

^aChromosomal abnormalities were detected by SKY analysis of mouse OHT− and OHT+ cell lines. Clonal chromosomal abnormalities are defined by a chromosome loss in at least three cells and a chromosome gain or a structural rearrangement in at least two cells. All numerical and structural changes are expressed are expressed in relation to the hypotetraploidy level (4n-). T, translocation (the chromosome listed first contains the centromere); Del, deletion; Dic, dicentric; TAs, possible telomeric association. Common abnormalities in OHT− and OHT+ cell lines are indicated in boldface.

FIG. 7.

Loss of PRMT1 in MEFs leads to genomic instability by SKY analysis. SKY was performed on the PRMT1^{FL/−;CreERT} MEF cells untreated (-OHT) or treated with OHT for 6 days (+OHT). SKY was carried out as described in Materials and Methods. A summary of all chromosomal abnormalities for each cell line is given in Table 1. Two representative metaphase spreads of the +OHT cell line analyzed by SKY are shown. (A) Spectral karyotype of a metaphase with 142 chromosomes, represented in classification pseudocolors. (B) Spectral karyotype of a metaphase with 71 chromosomes. The karyotype is classified with the inverted-DAPI image (left) and the classification pseudocolors (right). Arrows indicate the chromosomal abnormalities that are found only in the +OHT cell line, including unique chromosome deletions and translocations, chromosome breaks, and dicentric chromosomes. The inverted-DAPI image of this metaphase is shown in Fig. S2 in the supplemental material.

PRMT1 is required for both the G₂/M and the G₁/S checkpoint activation in response to IR-induced DNA damage.

In response to DNA damage, cell cycle checkpoints are activated to arrest cell cycle progression allowing time for repair. The G₁/S checkpoint prevents damaged DNA from being replicated and the G₂/M checkpoint prevents cells from entering mitosis with damaged DNA. We first analyzed the G₁/S checkpoint by measuring the abundance of cells in the S phase 20 h after IR treatment. In the absence of OHT, approximately 47 and 28% of PRMT1^{FL/−;CreERT} MEFs entered the S phase of the cell cycle with no treatment and 10 Gy of IR, respectively (Fig. 8A, −OHT). In contrast, 40 and 36% of the OHT-treated PRMT1^{FL/−;CreERT} MEFs entered the S phase after no treatment and 10 Gy of IR (Fig. 8A, +OHT). These findings show that PRMT1^−/− MEFs induced with OHT clearly have lost their S-phase checkpoint and ∼90% of the cells incorporated BrdU after DNA damage (Fig. 8A). An S-phase ratio was obtained for the −OHT- and +OHT-treated samples from two experiments in duplicate, and this ratio closely approaches 1 in PRMT1-deficient cells (+OHT, 0.8), whereas the −OHT MEFs had a ratio close to 0.5, and the difference was significant (P < 0.05) (Fig. 8A, lower panels).

FIG. 8.

PRMT1 is required for both the G₂/M and the G₁/S checkpoint activations in response to IR-induced DNA damage. (A) Duplicate cultures of PRMT1^{FL/−;CreERT} MEFs were left untreated (-OHT) or treated with OHT (+OHT). One set was treated with 10 Gy of IR (10 Gy IR), and the other set was left untreated (-IR). At 20 h after treatment, the cells were incubated with 10 μM BrdU for 1 h and stained with PI and an FITC-conjugated anti-BrdU antibody. The cells were analyzed by flow cytometry. A typical fluorescence-activated cell sorting profile is shown; the experiments were performed twice in duplicate, and the average is shown. The S-phase ratio was obtained by dividing the percentage of gated BrdU-positive cells treated with 10 Gy of IR by the percentage of gated BrdU-positive cells without IR treatment. Statistical significance was assessed by using Student t test. (B) Duplicate cultures of PRMT1^{FL/−;CreERT} MEFs were left untreated (-OHT) or treated with OHT (+OHT). One set was treated with 2 Gy of IR (2 Gy IR), and the other set was left untreated (-IR). At 1.5 h after treatment, the cells were fixed and stained with PI and anti-pSer10-histone H3 antibody to identify the cells in mitosis. The percentage of pSer10-histone H3-positive cells (circled area) was determined by flow cytometry. The experiments were performed twice in duplicate. A typical result showing anti-pSer10-histone H3 antibody staining and the M-phase ratio, derived by dividing the percentage gated pSer10-histone H3-positive cells treated with 2 or 10 Gy of IR by the average percentage of gated pSer10-histone H3-positive cells without IR treatment. Statistical significance was assessed by using the Student t test.

We then studied the contribution of PRMT1 to G₂/M checkpoint activation by measuring the abundance of cells entering mitosis 90 min after IR treatment using the anti-phosphorylated S10 histone H3 antibody. Without IR exposure, the frequency of mitotic cells was comparable in OHT-treated and nontreated PRMT1^{FL/−;CreERT} MEFs (Fig. 8B) (1.55% versus 1.42%). After 2 Gy of IR treatment, only 25% of the PRMT1^{FL/−;CreERT} MEFs progressed to the M phase (from 1.55% to 0.40%), a finding consistent with the majority of wild-type cells arresting before mitosis in the presence of DNA damage. However, 95% of the OHT-treated PRMT1^{FL/−;CreERT} MEFs progressed through the M phase (from 1.42% to 1.35%) (Fig. 8B), which is consistent with the cells having lost their G₂/M checkpoint. A mitosis ratio was obtained for the −OHT- and +OHT-treated samples from two experiments in duplicate, and this ratio closely approached 1 in PRMT1-deficient cells (+OHT, 0.9) treated with 2Gy of IR, whereas the −OHT MEFs had a ratio close to 0.3, and the difference was statistically significant (P < 0.001) (Fig. 8A, lower panels). The usage of 10 Gy of IR considerably reduced the quantities of cells that stained with antiphosphorylated S10 histone H3 antibody; nevertheless, the difference in the mitosis ratio was statistically significant between the −OHT- and +OHT-treated PRMT1^{FL/−;CreERT} MEFs (Fig. 8B, lower panel).

PRMT1-deficient cells are hypersensitive to etoposide treatment.

We next examined whether PRMT1-deficient cells were hypersensitive to DNA-damaging agents. We wanted to verify whether PRMT1-deficient cells were hypersensitive to the topoisomerase II inhibitor etoposide, which is known to induce DSBs. Since the PRMT1^−/− MEFs die within a week or so, we utilized U2OS cells transfected with PRMT1 siRNA to perform a colony formation assay. We reasoned that the PRMT1 deficiency would be required transiently to signal DNA damage, and thus we should not need cells that harbor a stable knockdown of PRMT1. Transient knockdown studies are efficient in the human osteosarcoma cell line U2OS, and these are often used to study the DDR. U2OS cells were transfected with control or PRMT1 siRNA, and the latter cells displayed a reduced level of PRMT1 by ∼95%, as observed by immunoblotting using α-tubulin as a loading control (Fig. 9A). The knockdown of PRMT1 in U2OS exhibited a reduced cell growth phenotype consistent with what was observed in MEFs (Fig. 9B).We observed that PRMT1 siRNA-treated U2OS had an increased sensitivity to etoposide-induced DNA damage compared to control siRNA-treated U2OS (Fig. 9C). While control transfected U2OS (siGFP) required a 1 h treatment of 1 μM etoposide to achieve 50% cell death (Fig. 9C, left panel), the siPRMT1 transfected U2OS required a dose lower than 0.5 μM to achieve 50% cell death (Fig. 9C, left panel). Similarly a shorter treatment time (∼25 min versus ∼50 min) was required to kill 50% of the siPRMT1-transfected U2OS cells with 5 μM etoposide (Fig. 9C, right panel). These findings show that PRMT1-deficient cells are hypersensitive to the DNA-damaging agent etoposide.

FIG. 9.

U2OS cells with downregulated PRMT1 expression are hypersensitive to etoposide treatment. (A) U2OS cells were transfected with siPRMT1 and control siGFP, and the loss of PRMT1 expression was verified by immunoblotting. (B) U2OS cells were plated at 3 × 10⁵ cells/10-cm tissue culture dish and transfected with either siGFP or siPRMT1. The cell growth was monitored by counting with a Beckman-Coulter counter (n = 4). (C) Two days after transfection, the cells were treated with trypsin and 400 cells were replated on 10-cm tissue culture dishes. The cells were treated next day with etoposide with the desired dosage and time period as indicated. The cells were then washed twice with PBS and grown in regular medium. The visible colonies containing >20 cells were counted at 10, 12, and 14 days after treatment, respectively, and normalized to 100% without the treatment. Ten fields were randomly selected, and the total number of the colonies in the 10 fields was calculated and expressed as a percentage.

Downregulated PRMT1 expression leads to reduced recruitment of RAD51 at sites of DNA damage induced by IR.

In proliferating mammalian cells, the major mode of DNA repair is HR (64). One of the key protein complexes is the recombinase protein RAD51, which is essential in repairing DNA DSBs by HR (5). We reasoned that the genomic instability in PRMT1-deficient cells may be caused by a lack or impaired HR. Using anti-RAD51 antibodies, we examined whether the RAD51 recombinase was recruited to DNA damage sites after IR treatment, as detected by focus formation by indirect immunofluorescence. U2OS transfected with PRMT1 siRNA displayed 53BP1 and RAD51 DNA damage foci without exogenous DNA damage, a finding consistent with PRMT1-deficient cells harboring spontaneous DNA damage (Fig. 10). These findings show that RAD51 was able to form foci in PRMT1-deficient cells. We next examined the efficiency by which 53BP1 and RAD51 formed foci in the presence of 10 Gy of IR. The number of cells with >5 foci were counted as a time course of recovery after IR treatment (Fig. 10). 53BP1 foci formed equally well in both siGFP- and siPRMT1-transfected U2OS cells, and the difference was not statistically significant (Fig. 10B, left panel). However, we observed that there was a statistical significant decrease in RAD51 foci that were formed at 2, 6, and 20 h after IR treatment (Fig. 10B, right panel). These findings show that the loss of PRMT1 results in the impairment of IR-induced RAD51 foci.

FIG. 10.

Downregulation of PRMT1 expression in U2OS cells leads to a reduced recruitment of RAD51 to DNA damage foci after IR treatment. U2OS cells were transfected with control siGFP and siPRMT1 and, 96 h after transfection, the cells were left untreated or treated with 10 Gy of IR. After various hours of recovery, the cells were visualized by indirect immunofluorescence with anti-53BP1 and anti-RAD51 antibodies. (A) Representative images. (B) Cells with >5 foci were counted and are expressed as a percentage. The graphs show the average and standard error of the mean from two independent experiments performed in duplicates where >10 different fields were analyzed. In total, more than 200 cells were counted for each sample in each experiment. Statistical significance was assessed by using the Student t test.

DISCUSSION

In the present study, we generated a conditional PRMT1 null allele in mice. Using the Cre/lox-conditional system, we show that the loss of PRMT1 expression leads to the loss of arginine methylation of substrates harboring a GAR motif, including Sam68 and MRE11. PRMT1-deficient cells display genomic instability and cell growth arrest. Furthermore, the cells exhibited spontaneous DNA damage, cell cycle progression delay, checkpoint defects, polyploidy, and chromosome instability. Using PRMT1 siRNA in U2OS cells, we further showed that PRMT1-deficient cells are hypersensitive to the DNA-damaging agent etoposide, and the cells exhibit a defect in IR-induced RAD51 recruitment at DNA damage foci. These data show that PRMT1 is required for genome maintenance and cell proliferation and highlight a key role for arginine methylation in the DDR pathway.

A PRMT1 hypomorphic allele generated by a gene-trapping strategy which maintains partial PRMT1 expression revealed the need for PRMT1 for early embryonic development as the embryos died at ∼E6.5 (44). ES cells derived from the PRMT1 hypomorphic allele harbor numerous hypomethylated proteins, including Sam68 and MRE11 (11, 20). However, these ES cells did not reveal the essential role for PRMT1 in genome maintenance and cell proliferation. Our findings that the loss of PRMT1 in MEFs results in genomic instability and polyploidy suggests that it may be the residual PRMT1 expression in ES cells that maintains cell viability. Alternatively, it is possible that somatic cells such as fibroblasts (i.e., MEFs) are more sensitive to the lack of arginine methylation by PRMT1. Saccharomyces cerevisiae contains one homolog of PRMT1, Hmt/Rmt1; yeasts null for this methyltransferase are viable (28) and exhibit the hypomethylation of RNA-binding proteins (50), and a role in the maintenance of silent chromatin was observed (69).

The role of arginine methylation, and therefore PRMTs, in cell cycle progression and checkpoint activation is not well characterized. The PRMTs are known to methylate histones and in a histone-dependent manner influence the cell cycle (7). PRMT1 has been shown to methylate H4R3 (52, 61) and cooperates with p300/CBP and CARM1 to activate gene expression (3). Leukemia cells containing the MLL-EEN fusion protein recruit PRMT1 in a Sam68-dependent manner to activate the HoxA9 gene (18). The knockdown of PRMT1 in this circumstance prevents leukemia cell growth (18). CARM1 is required for estrogen-induced cell cycle progression in MCF-7 breast cancer cells (26). The loss of PRMT5 using siRNA leads to the inhibition of cell growth of transformed B cells (62), while overexpression of PRMT5 in NIH 3T3 cells leads to the inhibition of cell proliferation (23). The phenotype we observe with the loss of PRMT1 may only be partly explained by H4R3 methylation, since the loss of PRMT1 changes the expression of only a few genes (36; data not shown).

The role of PRMTs in the DDR is also poorly characterized. We showed previously that mutation of the arginines within the GAR motif of MRE11 severely impairs its exonuclease activity but not its complex formation with NSB1 and RAD50 (11). The GAR motif expressed as a GFP fusion in mammalian cells was sufficient to target to DNA damage foci (21), suggesting that arginine methylation may regulate its interaction with DNA or with the recruitment of subsequent proteins for DNA repair (12, 21). We examined RAD51 foci since HR is dependent on the resection of DSBs by MRN (15) and its recruitment to the break should be impaired under these conditions. The reduced formation of RAD51 foci with IR treatment in PRMT1 siRNA-treated U2OS cells suggests that this defect may be contributed in part by the hypomethylation of MRE11. Another DDR protein that is methylated by PRMT1 is 53BP1, and its arginine methylation was shown to impact its ability to associate with DNA (13) and oligomerize (1). Although general methylase inhibitors prevent the formation of 53BP1 foci, the GAR motif is not required to localize 53BP1 to DNA damage sites (13, 46), since this property is dictated mainly by lysine methylated histones and the tandem Tudor domain of 53BP1 (30). Here, we extend these findings and show that the ability of 53BP1 to localize to DNA damage foci is not affected by the loss of PRMT1. The challenges that lie ahead will be to identify other PRMT1 substrate(s) required for genome maintenance and cell proliferation.

PRMT5, PRMT6, and PRMT7 also play a role during DNA damage. PRMT6 was shown to methylate polymerase β; however, the properties of PRMT6-deficient cells were not characterized (22). Interestingly, the reduced expression of PRMT7 sensitizes cancer cells to camptothecin (60), etoposide (8), and DNA-damaging agents (29). These findings suggest that symmetrical dimethylarginine methylation may also regulate the DDR. Recently, PRMT5 was shown to methylate p53 and regulate its ability to induce cell death and p53-dependent gene expression (32).

The molecular and cellular defects observed with the loss of PRMT1 mimic that of proteins involved in the DDR pathway. Hypomorphic alleles and animal models of the MRN complex display spontaneous DNA damage, hypersensitivity to DNA damage, checkpoint activation defects, and DNA repair defects (2, 15, 31, 34, 51, 57, 63). In proliferating cells, the loss of components required for HR repair, including ATR (14), MRE11 (15, 57), RAD50 (2), NBS1 (34, 63), and RAD51 (59), lead to genomic instability and cell death. Our findings that PRMT1-deficient cells have spontaneous DNA damage, have checkpoint defects, are hypersensitive to DNA-damaging agents, show chromosome instability, and are impaired in recruiting RAD51 to DNA damage sites induced by IR suggest that PRMT1 is a key player in the DDR pathway. Interestingly, the loss of PRMT1 induces polyploidy in the absence of apoptosis and this is similar to certain Myc mutants (24). Therefore, the loss of PRMT1 may have two separate activities: one related to the cell survival as recently reported (67) and the other via the induction of genomic instability. The result is a polyploid growth arrested cell with increased level of damaged genomic DNA. In PRMT1-deficient cells, we observed cell cycle delays that may suggest checkpoint activation. It remains to be determined whether the low doses of spontaneous DNA damage are responsible of this observation. Recently, it was shown that the loss of a component of the MRN complex, RAD50, in postmitotic tissues is dispensable (2). These findings suggest that the loss of PRMT1 may, like that of RAD50, be tolerated in postmitotic cells and suggest that PRMT1 may represent an important target for cancer treatment.